[Bf-python] After 2.28

[Michel Selten]

On Mon, 2003-07-21 at 03:09, Willian Padovani Germano wrote:
> That'll make us reach the first milestone: substitute the old
> implementation with a brand new one, written in C and much easier to
> read (and readability will probably improve, once we get time to
> rearrange a few things).

Yay! I'm very happy with the result so far! When I started this, I had
no idea what big a task lay ahead. Thanks to all who implemented parts
of the API! I couldn't have done it alone - at least not in this time
frame.

My reason for starting this rewrite are probably known to all, but I'm
willing to repeat. Back in January or February, we were having a lot of
trouble with the old implementation, it caused build problems, had some
difficult to understand macro magic and seemed difficult to be extended.
I took on the challenge to start it all over and after the first few
parts had been implemented, Willian joined in to help me out! Later on
Jacques and Jordi joined in too and wrote either new modules and/or
existing modules.

My initial goal was to have a clean, easy to read and easy extendable
implementation of the Python API. When Willian joined in, and very
quickly came with the first module he implemented, I already had the
feeling I was heading in the right direction. Now that my goals are met,
I want to focus on other parts of Blender development too.

Of course I will still be around for fixing bugs and possibly implement
a small function or so, but I won't be implementing any new big modules.
And I will follow the discussions on the Python mailing list - plus give
feedback if I think I have something sensible to say :)

I have asked Willian to take over the 'responsibility' as a module owner
for Python. And most of you probably know that he accepted - it has come
up in the various meeting minutes written by Ton.

I think the current developers working on improving Python are doing a
great job! And I trust they will be doing so in the future. There are a
lot of areas where things can be improved and extended.

> Now that the new code is "in", we have time to start discussing the next
> goals.

Okay, I'm willing to give some goals I'm willing to participate in
development first: :)
- Remove some of the duplicate code that has entered the source tree.
  This duplicate code is difficult to maintain.
- Fix bugs

Some of the goals that could be tackled in the future:
- Support for all datablock types. This is simply working on the same
  kind of function methods we've been doing in the past. Should not
  be a difficult one. Plus, this can be done incrementally!
- Expose internal blender functionality to the Python API. For
  example, think of the basic mathematics functions.
- Expose default meshes (such as a circle, plane, sphere, ...)

Willian mentioned three items that indeed need to be tackled in the
future too - and all of them are definately not the 'standard implement
a new datablock module' thing :)

With regards,
	Michel